Predicting University Course Bookings via Machine Learning Classification

Supervisor: Univ.-Prof. Dr. Radu Prodan
Author: Maria Makhov

Abstract

The MCI Internationale Hochschule GmbH seeks to improve its workflow pertaining to course bookings by using AI to predict which course will need which room, lab, or lecture hall for which timeslot. The program will build on already gathered data from the past year and use Machine Learning, mainly Random Forest Classification as a first base, to provide recommendations for the user to then accept or decline. The dataset is on the smaller side (800-1000 courses), which lends itself to making use of simpler ML algorithms, but this could be subject to change during the research and coding process, depending on what issues might emerge. The stakeholder plans to use this as an extension to the already existing administrative software for the university (called “Designer”), which additionally already hosts the data needed to train the algorithm.